You are viewing a single comment's thread. View all
22
Supermatmike on scored.co
1 year ago22 points(+0/-0/+22Score on mirror)3 children
what we call "AI" today is not something to be worried about, it's simply very complex algorithm code, it's not intelligent, it does not think. at a base level it's no different from Minecraft world generation.
We're no closer to AI today then we were in 1980.
I think it will actually prove to be a boon. rendering one's ability to find the truth in digital spaces an impossibility. soon no one will be able to tell fact from fiction on the internet. everyone will be forced to go back to real life.
> "it's simply very complex algorithm code, it's not intelligent, it does not think"
It's not a complex algorithm. It's statistics, built on top of a staggeringly large dataset. It essentially does a playback of the thinking people used when writing the the terabytes of texts that was turned into hundreds of billions of rules like "if word A appears after word B then word C is more likely to follow" that an LLM consists of. Deeplearning models in other domains have similar kinds of rules, just on tokens that aren't word-fragments.
It's not intelligent, and doesn't think, and doesn't learn or have any form of memory. It gives the illusion of having a conversation by being fed all your previous questions and its replies every time you write a new question.
Despite it not being thinking, it is useful enough to solve a large number of real-world problems.
1 year ago2 points(+0/-0/+2Score on mirror)1 child
It will never happen.
Sam Hyde made a great point about the robot ascendency. It will never happen because the second a robot or self driving car accidentally kills a nigger, it's all shut down.
Did you hear about the time DARPA or whatever built an AI-powered automated sentry, and a bunch of low IQ marines were charged with getting close to it? Pretty much every one of them did. One of them just used a big cardboard box. Someone else beat it simply by wearing a mask and walking on all fours. Even basic camouflage confused the AI.
Well, to touch further on that, I don't really think that's as big of a problem as you're making it out to be. Firstly, armor is heavy. Any armor worth a damn would render a robot like that slow as a snail to outright stationary, and even if they equipped it with some sort of sci-fi bullshit armor that doesn't weigh anything, light it on fire or hit it with explosives, especially a shaped charge, and that thing is gonna be rendered into scrap metal pretty damn quick.
If it's the advanced detection tech you're worried about, do remember a couple months ago when a couple Marines fooled a DARPA detection AI by hiding under a cardboard box Solid Snake style.
And then, the most important detail I think on why this isn't a problem: cost. Sending out an army of militarized robots like this would be a monumental expense, from research and development, making sure these things even work to begin with, to actually building them, and then keeping them working in field conditions, mud, rain, long periods of sun exposure, extreme temperatures and terrain conditions, fielding an army of these things would be a financial and logistical nightmare.
As much as I hate to say it, it would be cheaper and dare I say MUCH more effective to just gather up all the niggers in the big cities, slap an AK in their hands and send them at us.
We're no closer to AI today then we were in 1980.
I think it will actually prove to be a boon. rendering one's ability to find the truth in digital spaces an impossibility. soon no one will be able to tell fact from fiction on the internet. everyone will be forced to go back to real life.
It's not a complex algorithm. It's statistics, built on top of a staggeringly large dataset. It essentially does a playback of the thinking people used when writing the the terabytes of texts that was turned into hundreds of billions of rules like "if word A appears after word B then word C is more likely to follow" that an LLM consists of. Deeplearning models in other domains have similar kinds of rules, just on tokens that aren't word-fragments.
It's not intelligent, and doesn't think, and doesn't learn or have any form of memory. It gives the illusion of having a conversation by being fed all your previous questions and its replies every time you write a new question.
Despite it not being thinking, it is useful enough to solve a large number of real-world problems.
I don't exactly see how that's cause for concern.
Sam Hyde made a great point about the robot ascendency. It will never happen because the second a robot or self driving car accidentally kills a nigger, it's all shut down.
If it's the advanced detection tech you're worried about, do remember a couple months ago when a couple Marines fooled a DARPA detection AI by hiding under a cardboard box Solid Snake style.
And then, the most important detail I think on why this isn't a problem: cost. Sending out an army of militarized robots like this would be a monumental expense, from research and development, making sure these things even work to begin with, to actually building them, and then keeping them working in field conditions, mud, rain, long periods of sun exposure, extreme temperatures and terrain conditions, fielding an army of these things would be a financial and logistical nightmare.
As much as I hate to say it, it would be cheaper and dare I say MUCH more effective to just gather up all the niggers in the big cities, slap an AK in their hands and send them at us.